- Top of page
- Materials and Methods
- Disclosure Statement
To image hypoxia-inducible factor-1 (HIF-1)-active tumors, we previously developed a chimeric protein probe ([123/125I]IPOS) that is degraded in the same manner as HIF-1α under normoxic conditions. In the present study, we aim to show that the accumulation of radioiodinated POS reflects the expression of HIF-1. In vivo single-photon emission computed tomography (SPECT)/X-ray CT (CT) imaging, autoradiography, and double-fluorescent immunostaining for HIF-1α and pimonidazole (PIMO) were carried out 24 h after the injection of [125I]IPOS. Tumor metabolite analysis was also carried out. A tumor was clearly visualized by multi-pinhole, high-resolution SPECT/CT imaging with [125I]IPOS. The obtained images were in accordance with the corresponding autoradiograms and with the results of ex vivo biodistribution. A metabolite analysis revealed that 77% of the radioactivity was eluted in the macromolecular fraction, suggesting that the radioactivity mainly existed as [125I]IPOS in the tumors. Immunohistochemistry revealed that the HIF-1α-positive areas and PIMO-positive areas were not always identical, only some of the regions were positive for both markers. The areas showing [125I]IPOS accumulation were positively and significantly correlated with the HIF-1α-positive areas (R = 0.75, P < 0.0001). The correlation coefficient between [125I]IPOS-accumulated areas and HIF-1α-positive areas was significantly greater than that between the [125I]IPOS-accumulated areas and the PIMO-positive areas (P < 0.01). These findings indicate that [125I]IPOS accumulation reflects HIF-1 expression. Thus, [123/125I]IPOS can serve as a useful probe for the molecular imaging of HIF-1-active tumors. (Cancer Sci 2011; 102: 2090–2096)
The expansion of a tumor mass implies the development of cancer cells that are distant from blood vessels. Cells outgrow their vascular supply, leading to the lowering of oxygen partial pressure below physiologic levels (i.e., hypoxia). Hypoxia-inducible factor-1 (HIF-1) is a key transcriptional regulator in response to hypoxia.(1) Overexpression of HIF-1 has been indicated in over 70% of human cancers (e.g. breast, prostate, brain, lung, head and neck cancers) and their metastases compared to that of adjacent normal tissue.(2) Clinical studies associate the hypoxic status of a tumor with bad prognosis and resistance to chemotherapy and radiotherapy.(3) Furthermore, several HIF-1 inhibitors have been developed to inhibit tumor growth and to overcome resistance to therapies.(4) Thus, non-invasive imaging of HIF-1-active regions in a tumor provides useful information for qualitative cancer diagnosis, cancer therapy, and improving the effectiveness of HIF-1 inhibitors.
The expression of HIF-1 in tumors has been shown to be affected by the microenvironment in each tumor, resulting in heterogeneous HIF-1 tumor activity and expression.(5) Moreover, HIF-1 has been reported to be expressed even in a small tumor spheroid (<1 mm in diameter).(6) Therefore, a high spatial resolution device was required for the accurate imaging of HIF-1-expressed regions. Several techniques to improve spatial resolution, such as depth-of-interaction scintillators, image reconstruction using time-of-flight information, and multi-pinhole collimators, have recently been developed(7–10) and incorporated into commercially available PET and single-photon emission computed tomography (SPECT) systems. These modalities enable the detailed visualization of intratumoral HIF-1 expression in both animal models and humans.
Hypoxia-inducible factor-1 is composed of two subunits, HIF-1α and HIF-1β. The former is rapidly degraded by the ubiquitin–proteasome pathway under normoxic conditions, but it is stabilized under hypoxic conditions and regulates HIF-1 transcriptional activity. The proteasomal degradation of HIF-1α occurs through protein hydroxylation on proline residues in the oxygen-dependent degradation domain (ODD) of HIF-1α.(11,12) Thus, it is expected that probes containing the ODD and that are degraded in manner similar to HIF-1α can evaluate HIF-1 activity in vivo.
Based on the above concept, we recently developed a protein in which the protein transduction domain (PTD)(13) and a monomeric streptavidin (SAV) are fused to the ODD. We also synthesized a radiolabeled biotin derivative, (3-[123/125I]iodobenzoyl)norbiotinamide ([123/125I]IBB) and conjugated the chimeric protein, PTD-ODD-SAV (POS), and [123/125I]IBB to produce [123/125I]IBB-POS ([123/125I]IPOS). POS was degraded in an oxygen-dependent manner, and a clear tumor image was obtained by planar imaging at 24 h after the injection of [123I]IPOS. The intratumoral distribution of [125I]IPOS corresponded to the hypoxic areas. However, direct comparison between the area showing [125I]IPOS accumulation and HIF-1α expression was not carried out in the previous study.(14) Moreover, the 3D distribution of HIF-1-active regions in tumors has not been extensively evaluated in vivo. We then compared [125I]IPOS distribution to HIF-1 expression by autoradiographic and immunohistochemical studies in the present study. Moreover, we attempted to image the detailed intratumoral distribution of [125I]IPOS, which potentially reflects the heterogeneous expression of HIF-1, using a multi-pinhole, high-resolution small animal SPECT. The purpose of the present study is to reveal whether [123/125I]IPOS can be used to visualize HIF-1-active regions in a tumor.
- Top of page
- Materials and Methods
- Disclosure Statement
A number of hypoxia imaging probes, such as [18F]fluoromisonidazole ([18F]FMISO), 1-α-D-(5-deoxy-5-[18F]fluoroarabinofuranosyl)-2-nitroimidazole ([18F]FAZA), and [60/64Cu]Cu-diacetyl-bis(N4-methylthiosemicarbazone) ([60/64Cu]Cu-ATSM), have been developed and used in clinical studies.(20,21) These probes detect physically low oxygen partial pressure (<10 mmHg) and are useful for predicting the efficacy of radiotherapy.(22) In fact, clinical PET studies with [60Cu]Cu-ATSM have shown an inverse relationship between the tumor uptake of [60Cu]Cu-ATSM and the response to therapy in patients with lung and rectal carcinomas.(21) However, the mechanisms underlying hypoxic accumulation of these probes are not dependent on HIF-1 expression and further studies are necessary to decisively determine the role of them in identifying HIF-1-active regions in tumors.
The expression level of HIF-1α was reported to increase dramatically as the oxygen partial pressure decreased below 40 mmHg.(23) Furthermore, it was reported that the oxygen concentrations at which HIF-1 activity occurs differ among tissue cells.(3,23) For example, pulmonary cells are consistently exposed to relatively high oxygen concentrations, and thus, the oxygen concentrations that these cells perceive as abnormal are relatively high. However, bone marrow cells, which usually exist under low oxygen concentrations, require no HIF-1 activity when they are kept under the same oxygen concentrations as pulmonary cells.(5) Therefore, probes that detect absolute physical oxygen partial pressures are not enough when attempting to detect the expression or activity of HIF-1. Because HIF-1 is associated with tumor progression, angiogenesis, metastasis, and resistance to chemo- and radiotherapy, HIF-1 imaging would provide significant information for characterizing tumor phenotype and would be useful for predicting prognosis. Thus, multiple details on tumor characterization, treatment strategy, and prognosis would be provided by linking [123I]IPOS-SPECT with [18F]FMISO-, [18F]FAZA-, or [60/64Cu]Cu-ATSM-PET.
Recently, nuclear medical imaging(24–26) and bioluminescent imaging(27,28) based on HIF-1-responsive reporter systems have been successfully carried out. However, these systems required exogenous gene transfection. To surmount this limitation, our research group developed a number of injectable protein probes, including [125I]IPOS, which contain ODD and degrade in an oxygen-dependent manner. These probes have been used successfully to visualize hypoxic tumors on planar nuclear medical imaging(14,15) and optical imaging.(1,5,29) However, the in vivo 3D distribution of HIF-1-active regions in tumors has not been extensively evaluated in our previous studies. Thus, we were challenged to obtain high spatial resolution tomographic images using multi-pinhole SPECT. Multi-pinhole SPECT systems have recently provided greatly improved efficiency with high spatial resolution and have overcome the resolution versus sensitivity trade-off.(30) The disadvantage of SPECT is a quantitative performance due to difficulties in attenuation and scattering correction, and the sensitivity of multi-pinhole SPECT is still lower than PET. However, because SPECT can theoretically achieve a higher spatial resolution than PET, wherein the resolution is limited by the positron range, it is suitable for small animal imaging. In fact, several multi-pinhole SPECT systems are able to image small components within mouse organs.(31,32) The in vivo images obtained in the present study were consistent with the results of ex vivo biodistribution. The tumor was clearly visualized, but no accumulation of radioactivity was observed in the opposite leg (Fig. 1A,B,D), indicating a high tumor/muscle ratio (10 in ex vivo biodistribution). Our observation of high radioactivity in the bladder (Fig. 1E) was in accordance with the finding that the radioactivity was mainly excreted in urine after the administration of [125I]IPOS.
The characterization of [125I]IPOS has been reported in a previous study.(14) Although the tumor uptake of [125I]IPOS was significantly and positively correlated with HIF-1 transcriptional activity, suggesting the feasibility of [125I]IPOS as an HIF-1-active tumor imaging probe, a direct comparison between the intratumoral distribution of [125I]IPOS and HIF-1 expression has not yet been carried out. In the present study, the majority of the [125I]IPOS-distributed areas in the tumor corresponded to HIF-1α-positive hypoxic areas, although there were a few unmatched regions (Fig. 3). A significant positive correlation was observed between the areas, and the correlation coefficient between [125I]IPOS and HIF-1α was significantly greater than that between [125I]IPOS and PIMO (Fig. 4). Although we cannot exclude the effect of the difference in the molecular sizes of [125I]IPOS and PIMO on the distribution, these findings are consistent with the results of recent studies showing that HIF-1α-positive areas are not always colocalized with PIMO-positive areas.(33–35) Thus, [125I]IPOS can be viewed as a probe that depicts HIF-1-active areas rather than the physically hypoxic areas themselves.
Although the detailed mechanism of cellular uptake and retention of [125I]IPOS is now under investigation, we found that low temperature (4°C) or ATP depletion inhibited the cellular uptake of [125I]IPOS (unpublished data). As most studies have shown endocytosis as the major cellular uptake pathway for most PTDs(36,37) and endocytosis occurs in an energy-dependent manner, [125I]IPOS is probably uptaken by the cells through the endocytotic pathway. After uptake by endocytosis, [125I]IPOS could be released from the endosomes and retained intracellularly because of its molecular size. In fact, in our previous experiment, [125I]IPOS was not cleared from the cells under hypoxic conditions for at least 24 h, after the incubation medium was replaced with fresh medium.(14) Although the intracellular radioactivity decreased gradually under normoxic conditions, the cleared radioactivity was mainly attributable to the small molecules.(14) These findings indicated the intracellular retention ability of [125I]IPOS.
In the present study, the biodistribution of [125I]IPOS was consistent with a previous report;(14) thus, the reproducibility of the biodistribution of [125I]IPOS was confirmed. Although the amount of tumoral accumulation of [125I]IPOS was not as high as that of 2-[18F]fluoro-2-deoxy-d-glucose ([18F]FDG), this could have been due to the limited expression of HIF-1. The expression of HIF-1 is not ubiquitous, but heterogeneous, and is small in tumors.(6) In fact, the probe accumulation in the tumors harboring HIF-1-dependent reporter genes was 1–2%ID/g in previous studies,(6,38,39) which is comparable to the results of the present study. The size-exclusion analysis revealed that the radioactivity in the tumor existed as 34-kDa macromolecules that were most likely [125I]IPOS existing stably in hypoxic regions. Furthermore, the radioactivity in urine and feces were mainly in low-molecular-weight compounds. This was consistent with our previous in vitro findings showing that [125I]IPOS was degraded in normoxic cells and that [125I]IBB and other small molecules were cleared from the cells.(14) Thus, these findings prove the concept of [125I]IPOS in vivo and show that [125I]IPOS exists stably only in the HIF-1-active regions while undergoing degradation and clearance in normoxic regions.
Although HIF-1α was reported to degrade within a few minutes under normoxic conditions,(5) the degradation of [125I]IPOS was relatively slow.(14) Thus, a long exposure time is necessary to obtain sufficiently contrasting images; this is one of the drawbacks of [125I]IPOS. Moreover, a study has shown that the acute hypoxic region created by the shutdown of blood vessels can be reoxygenated after reopening the blood vessels.(40) In such reoxygenated areas especially, the inconsistency in the rates of degradation of HIF-1α and [125I]IPOS may be responsible for the differences in the distribution pattern of [125I]IPOS and HIF-1α expression. POS comprises not only the ODD, which is the essential domain related to oxygen-dependent degradation, but also the PTD and SAV. This modification may hamper the degradation of the probe. However, this is not a substantial problem; a pretargeting approach can overcome this drawback. In fact, a 4–8-fold earlier visualization of the tumor can be achieved with the SAV–radiolabeled biotin-based pretargeting approach.(15,18)
In conclusion, multi-pinhole high-resolution SPECT/CT imaging with [125I]IPOS clearly visualized a tumor. The obtained images were in accordance with the corresponding autoradiograms and the analysis of ex vivo biodistribution. The autoradiograms documented that the intratumoral distribution of [125I]IPOS was heterogeneous and showed a good correlation with HIF-1α-positive areas. These findings indicate that the distribution of [125I]IPOS reflects the expression of HIF-1α, thus rendering [123/125I]IPOS an effective probe for the molecular imaging of HIF-1-active tumors.